\begin{table}
\centering
\caption{Impact of promotion on fuel collection and use in matched NGO/non-NGO villages}
\begin{adjustbox}{max width=\textwidth}
\label{t:impacts}
\begin{threeparttable}
\begin{tabular}{lcccc}
\toprule
& (1) & (2) & (3) & (4) \\ \cmidrule(lr){2-3} \cmidrule(lr){4-5}
& \multicolumn{2}{c}{Fuelwood use} & \multicolumn{2}{c}{Fuel-collection time} \\
& \multicolumn{2}{c}{(kilograms per day)} & \multicolumn{2}{c}{(minutes per day)} \\
\midrule
\({POST}_1\)&        4.42***&        3.38***&        3.96   &       -19.0   \\
            &      (0.68)   &      (0.57)   &      (18.4)   &      (18.5)   \\ \addlinespace
\({POST}_2\)&        1.50** &        1.44   &       -9.83   &       -30.7** \\
            &      (0.75)   &      (1.18)   &      (15.5)   &      (12.5)   \\ \addlinespace
\({TREATMENT}_j \times {POST}_1\)&       -2.01** &       -0.61   &       -25.6   &        19.2   \\
            &      (0.93)   &      (0.91)   &      (21.9)   &      (25.4)   \\ \addlinespace
\({TREATMENT}_j \times {POST}_2\)&        0.65   &        0.47   &       -14.0   &        12.8   \\
            &      (1.08)   &      (1.54)   &      (17.0)   &      (15.9)   \\ \addlinespace
\({NGO}_j \times {POST}_1\)&               &        2.24** &               &        53.3   \\
            &               &      (1.12)   &               &      (33.2)   \\ \addlinespace
\({NGO}_j \times {POST}_2\)&               &        0.12   &               &        48.5*  \\
            &               &      (1.47)   &               &      (29.0)   \\ \addlinespace
\({TREATMENT}_j \times {NGO}_j \times {POST}_1\)&               &       -2.93*  &               &       -95.1** \\
            &               &      (1.67)   &               &      (40.5)   \\ \addlinespace
\({TREATMENT}_j \times {NGO}_j \times {POST}_2\)&               &        0.34   &               &       -60.0*  \\
            &               &      (2.12)   &               &      (32.2)   \\
\midrule
Baseline non-NGO control mean & 8.9 & 8.9 & 114 & 114 \\
Household fixed-effects & Yes & Yes & Yes & Yes \\
\(N\)       &       1,143   &       1,143   &       2,829   &       2,829   \\
Adjusted \(R^2\)&        0.19   &        0.20   &       0.010   &       0.028   \\
\bottomrule
\end{tabular}
\begin{tablenotes}
{\setlength\labelsep{0pt}
\footnotesize
\item \textit{Notes}. The outcome variables are daily fuelwood use (columns 1--2) and total time spent collecting fuels (columns 3--4) as measured during baseline, and midline and endline follow-up surveys (conducted approximately three and fifteen months after the start of the intervention, respectively). Daily fuelwood use is derived from a 24-hour before--after household-level fuel-weighing test; we restrict the analysis in columns (1) and (2) to the subsample of households that participated in such tests in at least two of our three survey rounds \( (N = 388) \). The outcome variable for fuel-collection time is derived from self-reported data on time spent (per day, week, or month) collecting fuelwood, crop residue, leaves, dung, biomass pellets, kerosene, LPG, biogas, and---if relevant---any other fuel used by the household; missing observations for time spent collecting fuel for up to six households in each survey round are replaced with the survey-round-specific sample mean value. The even-numbered columns present the results of estimating the triple-differences specification outlined in Equation \eqref{e:triple-diff}. The accompanying odd-numbered columns present results from estimating a double-differences specification without NGO-specific interactions for comparison. Standard errors (in parentheses) are clustered at the hamlet level. \sym{*} \(p<0.10\), \sym{**} \(p<0.05\), \sym{***} \(p<0.01\).}
\end{tablenotes}
\end{threeparttable}
\end{adjustbox}
\end{table}
